Use of Cardiac Troponin Assays in Hemodialysis Patients

Despite its high incidence in hemodialysis (HD) patients, diagnosing acute coronary syndromes (ACS) in this population are challenging. Compared to non-dialysis patients, HD patients experiencing acute myocardial infarction are less likely to exhibit typical chest pain (44.4% vs. 68.3%) or display ST elevation on electrocardiogram (ECG) (19.1% vs. 35.9%) [1]. Complications such as cardiac arrest or in-hospital death, however, are almost twice as common [1].

During a dialysis session, patients often complain about unspecific symptoms, which may also accompany acute myocardial ischemia, such as hypotension (28.7%) or nausea/vomiting (11.7%) [2], complicating the correct diagnosis. Almost all dialysis patients have cTnT concentrations above the 99th percentile at baseline (99%), compared to (11%) for hs-cTnI [3] in the non-dialysis population. The fact that studies establishing optimal cut-off levels for cardiac troponins (cTns) in the diagnosis of ACS frequently excluded dialysis patients adds to this complexity.

A recent systematic review on intradialytic cTn changes illustrated the inconsistency of the available data. Problematically, previous studies did neither include patients with acute myocardial ischemia, nor provide information on HD treatment characteristics, which potentially influence biomarker levels, nor follow guideline-recommended pre-specified time points of blood sampling for ACS workup [4].

Despite acknowledging the prevalence of elevated cTn levels, current ESC and AHA/ACC/ASE/CHEST/SAEM/SCCT/SCMR guidelines on management of ACS lack information on diagnostic data related to cTn in end-stage kidney disease patients on HD, and importantly, during ongoing treatment.

The Standardized Outcomes in Nephrology Group-Hemodialysis (SONG-HD) working group acknowledged this problem and recommended that, in HD patients with an initial cTn value above the 99th percentile URL, a rise and/or fall of more than 20% is suggestive of an ACS in addition to clinical criteria, but also state that, so far, there is insufficient evidence for standard dialysis to alter the concentrations of cTn, impacting the diagnosis of ACS [5].

In a previous randomized controlled trial, we explored intradialytic changes of cTns [6]. In this study, patients without evidence of ongoing cardiac ischemia were treated with different, routinely used HD modalities. Therefore, changes of cTn levels exclusively reflect the effect of dialysis. Changes in cTnT concentrations were significant using high-flux membranes, the most frequently encountered membrane type in routine practice, medium cut-off (MCO) membranes and hemodiafiltration treatment. Thereby, when heeding the 0 h/1 h algorithm of the 2023 ESC Guidelines, a 1 h delta of cTnT >5 ng/L would be missed during HD treatment in almost all patients with NSTE-ACS treated with routinely used modalities. Furthermore, a recommended delta of >20% of cTn [5] occurs without evidence of ACS during dialysis with MCO membranes or HDF after 1 h of treatment. In this study, there were no clear trends in cTnI levels, which may indicate limited usability in this context, possibly due to adsorption of cTnI to the dialyzer, interference of heparin with epitopes or phosphorylation [7].

This translates into the clinical scenario where myocardial ischemia is suspected, but due to dialysis, the intradialytic measurements of cTnT may not increase despite the presence of ongoing ischemia. Thus, a NSTE-ACS can be missed during dialysis because of removal of cTnT.

The problem clinicians are currently facing is underscored by the following exemplary case: consider a patient on HD due to end-stage kidney disease resulting from diabetic kidney disease, who experiences upper abdominal pain during dialysis. As diabetic patients on HD with myocardial ischemia may present with untypical symptoms, there should be heightened clinical suspicion for ACS, prompting the acquisition of an ECG and, moreover, cardiac biomarkers. Let us assume the patient’s initial cTnT value is 78 ng/L, which is a common finding for HD patients [1], and a second sample after 1 h shows a concentration of 80 ng/L, with no further change after 4 h of dialysis (Fig. 1).

In case the patient was dialyzed with a MCO membrane (yellow) or with hemodiafiltration (red) treatment, which have been shown to reduce cTnT levels by approximately 20% or 20 ng/L within 1 h [6], a stable level may suggest an increase in cTnT. In such cases, a diagnosis of non-ST-elevation myocardial infarction should be taken into consideration, and referral to cardiology is advisable. Conversely, if the patient undergoes dialysis with a low-flux hemodialyzer (blue), a small rise attributable to hemoconcentration could be expected and might not solely represent sufficient grounds for a referral to cardiology based on cTnT rise.

We agree that the final diagnosis of ACS is a clinical decision, incorporating baseline risk factors, clinical history, symptom type and location, response to medications (e.g., nitrates), new ischemic ECG changes, ultrasound findings of new wall motion abnormality, and cTn results. Non-STE-ACS without high-risk features is not typically a medical emergency which requires an immediate invasive strategy, thus there is no urgency to interrupt dialysis treatment immediately, given that coronary angiography is recommended within 24 h for these patients. However, we would like to emphasize that the absence of a rise in troponin levels during dialysis does not rule out ACS when certain dialyzers are used and therefore ambulatory patients, which do not undergo repeated testing for cTns, might be at high risk for missing the diagnosis. Furthermore, the large variability of cTnI found in previous studies, possibly due to adsorption to the membrane or interference of, e.g., heparin with the epitopes needs further clarification.

In summary, we highlight current issues and propose research recommendations which should be addressed in upcoming guidelines. First, it is essential to acknowledge the treatment characteristics of HD, such as the dialyzer used, when considering ACS during the procedure, urging nephrologists to communicate possible troponin changes to cardiologists based on the treatment modality. Both nephrologists and cardiologists need to be emphasized the impact of dialysis on troponin levels.

Furthermore, protocols need to be established, on recommendations when to initiate troponin sampling, which assay and which time points of sampling to be used, as well as usage of echocardiogram during dialysis in patients suspected having ACS. Ideally, a comprehensive clinical trial extending our previous work should be conducted, involving widely used membranes and treatment modalities, trying to establish cut-off values for cTnT and cTnI changes during treatment.